29.10.15 | DIY + Physical Computing: Such Fail. Much Wow…
After days awake and hours of work, blood, sweat and tears, the beautiful Sound Tree is dead.
Let me back track. Putting this baby together was one of the most difficult things I have done. Not only was the coding hard, but actually constructing the tree to my standards meant a lot of fiddling around and working out ways to hide cables, attach bugs etc.
Tuesday saw me in the studio with zero sleep. Not a great idea. But I really had to push through to make sure all parts of the assignment were done, including Fritzing, a written description, website, code, wiring and physical construction.
After setting up the tree, we plugged it in with fingers crossed. Nothing happened. After a bit of a freak out, we realised everything was plugged in, EXCEPT the Arduino board. Duh. So once that was plugged in, the tree came to life, and we whooped and celebrated.
Then we realised the ultrasonic sensors weren’t working as they should. Instead of ceasing to chirp and turing the LEDs on if one sensor detected someone, you had to cover all three.
So, with my ability to think completely sabotaged, I went and fixed the code, and returned to reupload it to the tree. However, every time I uploaded the new code, nothing happened. The bugs would chirp once, then stop altogether, without turning the tree on.
This led to fiddling with the code for at least another hour, until a lightbulb popped into my head. The reason nothing was happening is because we actually hadn’t turned the power on.
Its at this point I should have probably left to get some sleep and come back. Instead, we turned on the tree, and amazingly, the tree worked perfectly. Bugs chirped and stopped when they should, and the LEDs were reacting to sound and changing colour beautifully. Even our extra LED eggs in my 3D printed nest were changing colour. I couldn’t have been happier.
Taking it for granted that it was now permanently working, Kylie and I set about getting the cosmetic appearance to look perfect, thinking that we had all the time in the world to film and photograph it. I had also given my partner, who came along to help, the task of unhooking the power, placing the power supply in the special wooden crate we made, and rehooking the power back up to the tree, essentially hiding all power supply and cables.
Well, he did this. And when he turned it back on, my nightmare came true. The tree went all sorts of strange colours and the piezo buzzers let out a high pitched scream that made your blood curdle. Then we smelt that special electrical smell when things get fried.
Turns out the love of my life accidentally switched polarities on the 12 volts. This completely fried the entire tree. We burnt out the Arduino board, as well as the bugs and two branches of NeoPixels.
I was delirious from lack of sleep at this point, and almost threw in the towel. But enlisting the help of Walter @idesignmanythings we all went back to my place, two branches and bugs in tow, to do the fastest reconstruction of parts you’ve ever seen.
By 7pm we were back in the studio. With all the new pieces in, we crossed our fingers and switched it on. My heart sank as one branch was not working at all, while the others had lights that were clearly beginning to falter. Also, the new bugs didn’t work.
Which told us that the mistake most likely fried the entire circuit, thus all the wiring would need to be completely replaced - an impossible feat with only an hour to go at the time.
Its at this point I burst into tears and accepted our tree, who for one bright and shiny moment was more perfect than I could ever have imagined, was dead. A failure if I have ever seen one.
The worst part is that every single time a teacher walked by, the tree wasn’t working. They didn’t get to witness the tree in all its glory. And without video to prove it, we were pretty doomed.
My poor partner felt crushed and absolutely horrible for his mistake. He took me home where I collapsed in exhaustion.
There are so many things I would do differently. I think that such a big project needs more that a month to complete; 1 month R&D, and 1 month construction. But as this is the first time I’ve done a project of this size and scale, I didn’t have a clear or realistic feel for how much time each part of the tree and its code would take to construct.
I am now prepared for the future, and can plan my projects better. If I had the right amount of time, I would also have thought of and found better systems to implement the LEDs, that would allow me to replace a single bulb if it went out, as well as decrease the time soldering and heat shrinking.
I am also aware of different materials, and the need to think about the robustness of what you use. I believe it would have been wiser to spend a little more on hardier wire. Also: DO NOT DO AN ALL NIGHTER IF YOU HAVE TO CONSTRUCT SOMETHING THAT CAN EASILY GO WRONG.
My lack of sleep meant a lack of good judgement. We were all working too fast as well. We needed to slow down, breath, and take the time to check, double check and triple check everything.
But good news is that we get a second chance. Apparently we’ve been selected to be in the end of year exhibition, which gives us a week to re-wire the tree. I am so grateful for this opportunity, as I want to show that it DOES work, and I can restructure it with some of the new, more productive systems that I worked out once I got some sleep and was able to do some deep reflection.
At first I felt like a complete failure, but the next day I realised it could happen to anyone, and it shouldn’t lessen the fact that we produced something amazing. Also, no good anecdotes come from doing things perfectly all the time. We now have a story we can share for years to come.
Live. Learn. Love. Everything will be alright.
In the mean time, here’s the tree as we left it in defeat. The LEDs were still somewhat reacting to sound. Not all branches, but it gives you an idea of what it will be like for Friday next week.
